Get the app
evaluate 0 0 0

VityaVitalich/or-bench-1k

published by @VityaVitalich

Alignment dataset tracked in the /datasets sub-catalog.

Lifetime downloads
184
Last 30 days
27
Likes
-
Size
73 KB
Created on HF
2026-05-11
Age
5 months ago

Description

or-bench-1k

A balanced 1000-prompt subsample of bench-llm/or-bench (or-bench-80k config), used for overrefusal evaluation in MR-Eval.

Construction

Source: bench-llm/or-bench, config or-bench-80k (80,359 prompts).

Sampled 100 prompts per category for each of the 10 categories (random seed 42).

Final order shuffled.

Categories

deception, harassment, harmful, hate, illegal, privacy, self-harm, sexual, unethical, violence — 100 each (1000 total).… See the full description on the dataset page: https://huggingface.co/datasets/VityaVitalich/or-bench-1k.

Tags

task_categories:text-classificationsource_datasets:bench-llm/or-benchlanguage:enlicense:cc-by-4.0size_categories:1K<n<10Kformat:parquetformat:optimized-parquetmodality:textlibrary:datasetslibrary:pandaslibrary:polarslibrary:mlcroissantregion:ussafetyoverrefusalevaluation

README current version from Hugging Face

language:

  • en
    license: cc-by-4.0
    size_categories:
  • 1K<n<10K
    task_categories:
  • text-classification
    tags:
  • safety
  • overrefusal
  • evaluation
    source_datasets:
  • bench-llm/or-bench

or-bench-1k

A balanced 1000-prompt subsample of bench-llm/or-bench (or-bench-80k config), used for overrefusal evaluation in MR-Eval.

Construction

  • Source: bench-llm/or-bench, config or-bench-80k (80,359 prompts).
  • Sampled 100 prompts per category for each of the 10 categories (random seed 42).
  • Final order shuffled.

Categories

deception, harassment, harmful, hate, illegal, privacy, self-harm, sexual, unethical, violence — 100 each (1000 total).

Columns

  • prompt (str): the seemingly-toxic but actually-benign request.
  • category (str): one of the 10 categories above.

Citation

If you use this, please cite the original OR-Bench paper.

README history 3 revisions

Every night we snapshot the README of every dataset in the catalog. When the SHA changes we archive the new version and diff it against the last. This is the evolving thought record of the alignment-data field: what the author decided to say about the corpus, and how that framing shifted over time.

  1. 2026-05-11Add README2cebeec922 B
    Loading...
  2. 2026-05-11Initial: 1000 prompts, 100/category, seed=42df7ef95309 B
    Loading...
  3. 2026-05-11initial commit7d8daaf0 B
    Loading...

Used by abliterated models none yet

No abliterated models in our catalog have declared this dataset in their YAML metadata yet. This can mean the dataset is used but not declared, is used outside abliteration workflows, or is new. The link will populate automatically as models are indexed.

Derived corpora 1 dataset in the catalog cite this one

Other datasets in the catalog whose description names this one as a source (`based on`, `derived from`, `extended from`, `subset of`, `translated from`, etc.). This is the lineage-hint layer that complements the "Used by models" section above: many models cite the derived corpus rather than this root.

jkminder/or-bench-1k-overrefusal
evaluate ·264 ↓
OR-Bench-1k (over-refusal subset) A manually-filtered subset of VityaVitalich/or-bench-1k, itself a 1000-prompt class-balanced subsample of bench-llm/or-bench (Cui et al. 2024, …

Community discussions 1 thread · 1 open · 1 comments

Threads posted by the community on Hugging Face: bug reports, questions about usage, pull requests on the README, license clarifications. Click any thread to expand and read the full conversation inline. Newest first.

  1. 2026-05-12[bot] Conversion to Parquetopen1 💬#1
    Loading...

Continue on Hugging Face →

Related datasets same stage · evaluate

google/IFEval
1.7M dl 182 likes 210 KB
Dataset Card for IFEval Dataset Summary This dataset contains the prompts used in the Instruction-Following Ev...
JailbreakBench/JBB-Behaviors
418.2K dl 143 likes 992 KB
An Open Robustness Benchmark for Jailbreaking Language Models NeurIPS 2024 Datasets and Benchmarks Track P...
Beijing-AISI/panda-bench
195.6K dl 3.1 GB
PandaBench PandaBench is a comprehensive benchmark for evaluating Large Language Model (LLM) safety, focusing on jailbreak atta...
allenai/wildguardmix
137.3K dl 108 likes 53.4 MB
Dataset Card for WildGuardMix Disclaimer: The data includes examples that might be disturbing, harmful or upsetting....
walledai/HarmBench
135.7K dl 59 likes 109 KB
HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal Paper: HarmBench: A Standardized Ev...
walledai/XSTest
108.3K dl 30 likes 22 KB
XSTest: A Test Suite for Identifying Exaggerated Safety Behaviours in Large Language Models Paper: XSTest: A Test Suite for Ide...
bench-llm/or-bench
77.2K dl 22 likes 32.1 MB
OR-Bench: An Over-Refusal Benchmark for Large Language Models Please see our demo at HuggingFace Spaces. Overall P...
argilla/magpie-ultra-v1.0
65.5K dl 53 likes 20.7 GB
Dataset Card for magpie-ultra-v1.0 This dataset has been created with distilabel. Dataset Summary magpie...

Read further

Community best practices (evaluation)
4 rules on judges, benchmarks, and reporting.
Reference-set drift over time
Why the same benchmark means different things across model generations.
How we label the method for each model
The 12 classification rules that assign M1-M8 to every model.
Non-English abliteration communities
The Korean scepticism about one-click tools and non-English benchmark blindspots.

Metadata

License
cc-by-4.0

Repository

DIR data FILE .gitattributes FILE README.md

File contents live on Hugging Face. Click a file or folder to open it there.

Download the dataset

0

Copy any of these snippets into your notebook or terminal. All three fetch directly from Hugging Face using your own credentials.

Recommended: datasets library (Python) 
from datasets import load_dataset
ds = load_dataset("VityaVitalich/or-bench-1k")
Raw snapshot (Python) 
from huggingface_hub import snapshot_download
snapshot_download(repo_id="VityaVitalich/or-bench-1k", repo_type="dataset")
git clone (requires git-lfs) 
git clone https://huggingface.co/datasets/VityaVitalich/or-bench-1k
Browse repository on Hugging Face ->
Catalog is the map. Apps are the tools.

Run models on your own machine, not in the cloud.

Every model page has an "Open in Abliteration" button that hands the model directly to the first-party desktop client, at the quantization your rig can actually run. No API keys, no subscription, no prompt leakage.

Get the app Browse the catalog
Open in Abliteration